deepspeedai / deepspeedai/DeepSpeed

[REQUEST] Saving model weights only in checkpoints

Open
#3,820 1 comment 7 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
43.1k
Forks
5k
Avg merge
4d 15h
Merged PRs (30d)
112

Description

Currently, in ZeRO-3, we are saving model params and optimizer states together in .*_optim_state.pt files. Saving optimzier states may greatly increase the checkpoint size, while we don't actually needs them for inference.

In order to extract the model weights, we need to load all checkpoint shards into memory, then extract the weights out, which requires a huge amount of RAM.

Therefore, it would be great to enable only saving model weights in checkpoints.

### Tasks

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by tracing the ZeRO-3 checkpoint save path that produces *_optim_state.pt files and identify where model parameters and optimizer states are combined. Done means checkpoints can contain only model weights, and extracting those weights no longer requires loading optimizer states or all shards into memory.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
distributed-systems, machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.